You are on page 1of 13

39

Aspen Tutorial #5: Sensitivity Analysis and Transport


Properties
Outline:
Problem Description
Updating the Simulation
Sensitivity Analysis
Transport Properties
Problem Description:
A mixture containing 50.0 wt% acetone and 50.0 wt% water is to be separated into two
streams one enriched in acetone and the other in water. The separation process consists
of extraction of the acetone from the water into methyl isobutyl ketone (MIBK), which
dissolves acetone but is nearly immiscible with water. The overall goal of this problem is
to separate the feed stream into two streams which have greater than 90% purity of water
and acetone respectively.
Up to this point we have not maximized our use of Aspens computational abilities.
Often times in chemical engineering we are faced with problems that have iterative
solutions or iterative steps on the way to a desired result (i.e. purity of a component in a
separation process based on a feed of another). This week we will be using Aspen to
calculate the flow rate of a second feed stream of MIBK, in order to get the desired >90%
purity of our water stream through the use of a sensitivity analysis. During a sensitivity
analysis (or design specification) Aspen iterates its calculation sequence through a range
of values provided for an independent variable, in order to obtain a specified result for a
dependent variable (within a certain tolerance).
Updating the Simulation:
The most realistic separation results that we obtained last week were based on using the
NRTL thermodynamic method. Make sure your simulation is set to this base method and
then reinitialize your simulation.
Add a second mixer and a second flash separation unit to your process flowsheet and
name them as you see fit. Connect the stream that is primarily water and acetone (the
stream off of the bottom of the first flash separator) to the new mixer and add in a new
feed stream of MIBK that also feeds into this new mixer. Next, connect the product from
this mixer to the new flash separation unit and add in the required product streams. Your
process flowsheet should now look like that seen in Figure 1.

Aspen Tutorial #5
40

Figure 1: Updated Process Flowsheet
Now open up the Data Browser window to update the inputs for the new additions to
your process flowsheet. The new feed stream of MIBK should have a flow rate of 50
lbs/hr of pure MIBK at a temperature of 75 F and a pressure of 50 psi. The new mixer
and flash separation units should be specified to be at 75 F and 50 psi.
If you run the simulation at this point, you should get results similar to those seen in the
stream table shown in Figure 2. You will notice that we do not get the desired 90%
purity of the water stream that is specified in the original problem description. While we
could simply rerun the simulation a few times to determine a feed rate of MIBK that
would give us this desired purity, we will instead program Aspen to complete the
iterations for us before reporting the results.
You may notice that the stream table shown in Figure 2 does not include all of the
streams. You might remember that this was discussed in Tutorial #2 under the Display
Options. I have shown only the important feed and product streams to save space (I have
eliminated all of the intermediate streams and the product streams with no flow).
Aspen Tutorial #5
41
Tutorial 5 - Sensitivity Analysis
Stream ID FEED MI BK1 MI BK2 M-A1 M-A2 WATER
Temperature F 75.0 75.0 75.0 75.0 75.0 75.0
Pressure psi 50.00 50.00 50.00 50.00 50.00 50.00
Vapor Frac 0.000 0.000 0.000 0.000 0.000 0.000
MoleFlow lbmol/hr 3.636 0.998 0.499 1.938 0.725 2.470
Mass Flow lb/hr 100.000 100.000 50.000 141.052 59.825 49.123
Volume Flow cuft/hr 1.825 2.009 1.004 2.772 1.181 0.818
Enthalpy MMBtu/hr -0.435 -0.140 -0.070 -0.246 -0.096 -0.303
Mass Frac
WATER 0.500 0.041 0.027 0.868
ACETONE 0.500 0.263 0.127 0.108
METHY-01 1.000 1.000 0.697 0.846 0.023
MoleFlow lbmol/hr
WATER 2.775 0.319 0.089 2.367
ACETONE 0.861 0.638 0.131 0.092
METHY-01 0.998 0.499 0.981 0.505 0.012

Figure 2: Stream Results with 50 lbs/hr MIBK Feed
Sensitivity Analysis:
Select the Flowsheeting Options tab in the Data Browser window and open up the Design
Spec option. At the bottom of the screen, select the new button and choose a name for
this design specification. When you have done this the Data Browser window should
look like that seen in Figure 3. You will notice that there are three areas where we must
input data in order for the required input to be complete. These are the tabs Define, Spec,
and Vary.
In the Define tab the user must set the dependent variable that they are interested in. For
our case, this is the purity of the water product stream (or mass fraction of water). Select
new at the bottom of this screen and name the new variable WATER. After hitting OK,
the Variable Definition window will appear. In this window we need to specify that we
want our variable to be the mass fraction of water in the pure water product stream. In
the type box, select MASS-FRAC (you may want to note the many types of design
specifications one can specify by scrolling through the options in the type box at this
time). In the stream box that then appears, select your water product stream and under
the component box, select WATER. At this point your Variable Definition window
should look similar to that seen in Figure 4. The only difference should be in the stream
name, unless you have used the same stream names I have in your process flowsheet. Hit
the close button when you have completed this.
Aspen Tutorial #5
42

Figure 3: Design Specification Window

Figure 4: Completed Variable Definition Window
For our purposes we are now done inputting information into the Define tab and can
move on to the Spec tab. You will notice that we have three values we must input into
Aspen Tutorial #5
43
this window. The first, Spec, is the dependent variable that we want to set a target value
for. This is the variable that we just defined in the Define tab as WATER. Type this into
this box. Target is the numeric value that we would like our dependent variable to be
equal to at the completion of the calculation iterations. Our target value is 90%, or 0.90.
Finally, Tolerance is how close the solution determined by Aspen must be to our target
value before it is deemed acceptable. For our purposes, a tolerance of 0.1% is acceptable
(this is input as 0.001). After inputting this, the Spec window should look like that seen
in Figure 5.

Figure 5: Completed Spec Window
To complete the input for our sensitivity analysis, we must input which variable is to be
varied. This is done under the Vary tab. In this simulation, we are varying the flow rate
of MIBK in the second feed stream of MIBK (mine is entitled MIBK2). This is the
stream we just added to our simulation. Under the Vary tab select MASS-FLOW under
the type tab. Again, it is worth pointing out the many different variables that can be
manipulated in Aspen. Under stream, select the stream that corresponds to your second
feed stream of MIBK. Next, select METHY-01 from the components list. At this point
the Vary tab should look like that seen in Figure 6.
The values placed into the Manipulated Variable Limits boxes indicate the range that
Aspen can use during its iteration calculations. One thing to note is that the original input
value under the stream inputs must fall within the range that is input here. Remember our
original input was 50 lbs/hr. For this tutorial, input a variable range from 25-100 lbs/hr.
The other blocks that can be filled on this screen relate to the step size that Aspen takes
during its iteration calculations. It is not necessary for the user to input values into these
blocks, and we will use the default Aspen values.
Aspen Tutorial #5
44

Figure 6: Vary Tab Options
At this point, our required input should again be complete. The completed Vary tab is
shown below in Figure 7. We are now ready to run the simulation again and check its
convergence based on our input design specifications. Hit the run button at this time and
when the computer has finished its calculations, open up the Run Control Panel (see
Tutorial #2 for help with this).
The Run Control Panel indicates how many iterations Aspen made during its
determination of the flow rate that met our design specification. If completed correctly,
your simulation should have no warnings and no errors indicated in this window. You
will notice in Figure 8 that my simulation took 5 iterations to determine results that were
within the specified tolerance. We must also complete a cursory check of the simulation
results as discussed in Tutorial #2. This is especially important now that we have
introduced design specifications into the simulation. Close the Run Control Panel
window and open up the data browser to confirm that the simulation converged with
reasonable results.

Aspen Tutorial #5
45

Figure 7: Completed Vary Window
You will notice that the Convergence option under the Results Summary Tab in the Data
Browser window now has results. This window indicates the final value of the variable
and the error associated with this variable as shown in Figure 9. The Error column
indicates how far off the final dependent variable was from the specified value and the
Error / Tolerance column indicates how closely the design specification converged. A
value of 1 in this column means that the simulation barely converged while a value near 0
indicates excellent convergence.
The final place where the user can get information regarding the convergence of a
simulation is under the Convergence tab in the Data Browser window. In this window
one can actually see each of the values attempted by Aspen during its iteration cycle.
Aspen Tutorial #5
46

Figure 8: Run Control Panel

Figure 9: Convergence Results
Aspen Tutorial #5
47
Complete a cursory check of the other simulation results as discussed in Tutorial #2 and
if all of them look acceptable, proceed on to the next section.
Transport Properties:
Although we touched on some of the options for including selected physical properties in
stream tables, we did not touch on adding those properties that are important for mass
transfer (i.e. diffusivities). However, diffusivity is not one of the default variables that
are reported by Aspen and it is only reported if the user defines a specific property set.
The easiest way to do this is to modify an existing property set that reports other
parameters of interest and then have Aspen report this property set. Open up the Prop-
Sets option under the Properties tab in the Data Browser Window. Aspen has five default
property sets that can easily be added to a stream table. These five are summarized in
Table 1 below.
Table 1: Aspen Property Sets
Property Set Use Properties
HXDESIGN Heat Exchanger Design Thermal and Transport Properties
THERMAL Thermal Properties Enthalpy, Heat Capacity, Thermal Conductivity
TXPORT Transport Properties Density, Viscosity, Surface Tension
VLE VL Equilibrium Fugacity, Activity, Vapor Pressure
VLLE VLL Equilibrium Fugacity, Activity, Vapor Pressure
We will be modifying the TXPORT property set so that it includes diffusivity values for
our system. In the Prop-Sets window, select TXPORT and hit the edit button at the
bottom of the screen. The window that opens up is shown in Figure 10, on the next page.
Select the last box in the first column that is currently blank. In doing so, you will be
presented with a scrolling window of physical properties that Aspen can calculate for the
user. Scroll down until you find DMX, which is the variable for diffusivity in Aspen.
You will notice that a description of what each physical property is appears in the bottom
window as you scroll over the options. Aspen has seven built-in diffusivity models, some
of which you may be familiar with. These models are summarized in Table 2.
Table 2: Diffusivity Models
Model Equation Application
Chapman-Enskog-Wilke-Lee (Binary) Low Pressure Vapor
Chapman-Enskog-Wilke-Lee (Mixture) Low Pressure Vapor
Dawson-Khoury-Kobayashi (Binary) Vapor
Dawson-Khoury-Kobayashi (Mixture) Vapor
Nernst-Hartley Electrolyte
Wilke-Chang (Binary) Liquid
Wilke-Change (Mixture) Liquid
Aspen Tutorial #5
48

Figure 10: TXPORT Edit Window
Now select the Qualifiers tab. This window allows the user to input what phases they
would like the property set to be reported for. Because we are not concerned about the
vapor phase at this point, we will remove it from the reported results. Select the box
marked Vapor and hit the Delete key on the keyboard. The Qualifiers tab should now
look like that seen in Figure 11.
Aspen Tutorial #5
49

Figure 11: Qualifiers Window
We must now add the TXPORT property set to the stream table that is shown on the
process flowsheet. To do this we must go to the Report Options window under the Setup
tab in the Data Browser Window. Under the stream tab, hit the Property Sets button.
This will open up the window shown in Figure 12.

Figure 12: Property Sets Window
Select TXPORT and hit the single arrow button pointing to the right. This will move
TXPORT to the side labeled Selected Property Sets, and it will now be displayed in the
Aspen Tutorial #5
50
stream table. After you have done this, close the Property Sets window. To reduce the
number of variables shown in our stream table (to reduce its size), uncheck the mole flow
basis box. This will remove the mole flows from the stream table (all of our assigned
values have been mass flows so these have not played a role in our work yet). When you
have done this, reinitialize and rerun your simulation. In order to have the changes to the
stream table show up, you will most likely need to click on the stream table and then
click off of it. Another option is to delete the existing stream table and add a new one to
the process flowsheet. For comparison sake, my final stream table is shown below in
Figure 13. Unfortunately, the diffusivity values (with the units of ft
2
/hr) are too small to
show differences in the table. However, if you were to switch the units from the default
ones, you would get values that show differences in the three decimal places reported in
the table.
Tutorial 5 - Sensitivity Analysis
Stream ID FEED MIBK1 MIBK2 M-A1 M-A2 W-A1 WATER
Temperature F 75.0 75.0 75.0 75.0 75.0 75.0 75.0
Pressure psi 50.00 50.00 50.00 50.00 50.00 50.00 50.00
Vapor Frac 0.000 0.000 0.000 0.000 0.000 0.000 0.000
Mole Flow lbmol/hr 3.636 0.998 0.899 1.938 1.204 2.696 2.392
Mass Flow lb/hr 100.000 100.000 90.079 141.052 102.593 58.948 46.434
VolumeFlow cuft/hr 1.825 2.009 1.809 2.772 2.029 1.011 0.767
Enthalpy MMBtu/hr -0.435 -0.140 -0.127 -0.246 -0.161 -0.329 -0.293
Mass Frac
WATER 0.500 0.041 0.024 0.751 0.900
ACETONE 0.500 0.263 0.091 0.220 0.078
METHY -01 1.000 1.000 0.697 0.885 0.030 0.022
* ** LI QUI D PHASE ** *
Density lb/cuft 54.800 49.783 49.783 50.892 50.565 58.302 60.543
Viscosity cP 0.720 0.552 0.552 0.498 0.542 0.851 0.907
Surf aceTen dyne/cm 61.235 23.538 23.538 31.578 29.101 68.595 71.528
DMX sqft/hr
WATER <0.001 <0.001 <0.001 <0.001 <0.001
ACETONE <0.001 <0.001 <0.001 <0.001 <0.001
METHY -01 0.000 0.000 <0.001 <0.001 <0.001 <0.001

Figure 13: Final Stream Table



Next week: Separation Spreadsheets by Mark Burns, University of Michigan
Aspen Tutorial #5
51
Tutorial #5 Homework and Solution
Question:
What flow rate of MIBK is necessary to achieve 95% purity of the water stream? Show
your results with the stream table from your simulation. Hint: Modify your existing
design specification by changing both the target spec and the range for the independent
variable (I suggest an upper limit of 400 lbs/hr). If your upper limit is not increased
above the final result, your solution will not converge!
Solution:
From my Aspen simulation I obtained a feed rate of 324 lbs/hr MIBK, to get a water
purity of 95 wt%. This answer may vary between Aspen simulations, but your results
should be close to this (within 5 lbs/hr).

Tutorial 5- Sensitivity Analysis
StreamID FEED MIBK1 MIBK2 M-A1 M-A2 W-A1 WATER
Temperature F 75.0 75.0 75.0 75.0 75.0 75.0 75.0
Pressure psi 50.00 50.00 50.00 50.00 50.00 50.00 50.00
Vapor Frac 0.000 0.000 0.000 0.000 0.000 0.000 0.000
MoleFlow lbmol/hr 3.636 0.998 3.239 1.938 3.835 2.696 2.100
Mass Flow lb/hr 100.000 100.000 324.409 141.052 344.063 58.948 39.295
VolumeFlow cuft/hr 1.825 2.009 6.517 2.772 6.826 1.011 0.641
Enthalpy MMBtu/hr -0.435 -0.140 -0.456 -0.246 -0.525 -0.329 -0.258
Mass Frac
WATER 0.500 0.041 0.020 0.751 0.950
ACETONE 0.500 0.263 0.034 0.220 0.030
METHY-01 1.000 1.000 0.697 0.946 0.030 0.020
*** LIQUID PHASE ***
Density lb/cuft 54.800 49.783 49.783 50.892 50.407 58.302 61.333
Viscosity cP 0.720 0.552 0.552 0.498 0.564 0.851 0.924
SurfaceTen dyne/cm 61.235 23.538 23.538 31.578 28.470 68.595 72.376
DMX sqft/hr
WATER <0.001 <0.001 <0.001 <0.001 <0.001
ACETONE <0.001 <0.001 <0.001 <0.001 <0.001
METHY-01 0.000 0.000 <0.001 <0.001 <0.001 <0.001

You might also like